878 Inorganic Chemistry, Vol. 16, No. 4, 1977
Melvyn Rowen Churchill and Barry G. DeBoer
Contribution from the Departments of Chemistry, State University of New York at Buffalo, Buffalo, New York 14214, and the University of Illinois at Chicago Circle, Chicago, Illinois 60680
Structural Studies on Polynuclear Osmium Carbonyl Hydrides. 1. Crystal Structures of the Isomorphous Species HZOs3(CO)ll and O S ~ ( C O )Direct ~ ~ Information on the Role of an Equatorial p2-Bridging Hydride Ligand in Perturbing the Arrangement of Carbonyl Ligands in a Triangular Cluster MELVYN ROWEN CHURCHILL” and BARRY G. DeBOER2
Received July 2, 1976
AIC60462V
The complex H20S3(CO)ll,undecacarbonyldihydridotriosmiut 11, has been studied by a single-crystal x-ray structural analysis. The complex crystallizes in the centrosymmetric monoclinic pace group P2,/n with a = 8.0744 (16) A, b = 14.7265 (29) A, c = 14.7770 (28) A, B = 101.36 (l)’, and Z = 4. This species is found to be isomorphous with and quasi-isostructural to Os3(CO)12. The structure of O S ~ ( C Q[previously )~~ reported by Corey and Dahl in 19621 has been completely redetermined, accurate cell dimensions are a = 8.0817 (6) A, b = 14.7683 (11) A, c = 14.5773 (11) A, 0 = 100.56 (l)’, Z = 4, and space group P2Jn. Data for both complexes were collected with a Picker FACS-1 diffractometer (using Mo Ka radiation) and structural parameters were optimized via full-matrix least-squares refinement. Final discrepancy indices are RF = 3.68% and RwF= 3.52% for H20s3(CO)ll(2259 reflections) and RF = 3.35% and R w=~3.62% for O S ~ ( C O(3040 ) ~ ~reflections). The effects of the equatorial p2-bridging hydride ligand in H20s3(CO)11(relative to the isomorphous species O S ~ ( C O ) ~ ~ ) include a lengthening of the bridged osmium-osmium bond and angular redistribution of the equatorial carbonyl ligands.
+
+
hOl for h + I = 2n 1 and OM) for k = 2n 1. The centrosymmetric monoclinic space group P2,/n is thereby indicated. The crystal was We have been interested in the structural characterization transferred to a Picker FACS-1 automated diffractometer, was of species with bridging hydride ligands for the last 10 years accurately centered, and was oriented with [ 1011 offset by about 4 O and have reported structural studies of Z T ( B H ~ )HRez~,~ from coincidence with the q5 axis. Unit cell measurement and data collection were camed out as described previously;” details are given Mn(CO)14,4 [H2Re3(C0)i2-IIS [H6Re4(CO)iz2-]s6 Hzin Table I. RU6(co)i8? H6C%(PPh3)6,* [CSM~SR~CIIZ(C(-H)(C(-C~),~ The unit cell parameters obtained for H20s3(CQ),, were sufficiently [H2W2(CO)82-],’0H R U ~ ( C O ) ~ ~ ( C = N M ~and ~ ) , ”HOs3) ,= ~ 8.10 close to those reported by Corey and Dahl14 for O S ~ ( C O[a (CO)lo(CHCH2PMezPh).’2The overall structural effects of (3) A, b = 14.79 (4) A, c = 14.64 (4) A, B = 100” 27’ (20’); space bridging hydride ligands have been summarized in a recent group P21/n] to cause us considerable concern. However, Professor article.’Ib Shapley assured us that the sample of H z 0 ~ 3 ( C O ) Iwas I pure and We recently reported a preliminary account of the structure well-characterized and we found (later) that more precisely determined of H20s3(C0)1113 and noted that it was isomorphous with unit cell parameters for O S ~ ( C Owere ) ~ ~ significantly different from O S ~ ( C O )a~species ~, studied previously by Corey and Dahl.14 those of H20s3(CO)II.The two compounds are, nevertheless, both isomorphous and (as shown below) quasi-isostructural. We have now completed our structural study of H2Follow_ingdata collection, the H20~3(CO)ll crystal was zealignedO S ~ ( C O )and ~ ~ have also carried out an accurate new with [ 1011 along the q5 axis. Measurement of the axial 202 and 505 structural study of Os3(CO)12. Since the two species are reflections via f?-28 scans at loo intervals from q5 = 0 to q5 = 350’ isomorphous and close to isostructural, the differences in showed variations in intensity of 47.8% and 42.0%, respectively geometry should be attributable almost entirely to the presence [variation (%) = 100 X (maximum - minimum)/(average)]. These or absence of a p2-bridging hydride ligand. [The structures data were corrected for absorption along with the primary data set. of O S ~ ( C O and ) ~ ~H20s3(CO)l,are known to be I and 11, Variations were reduced to 13.5% and 9.4%, respectively. [These rather large residual variations arise principally as a result of the difficulty of measuring a highly absorbing small crystal with sufficient precision.] Data Collection for Os3(CO)l2. The crystal selected for data collection was multifaceted with approximate dimensions 0.093 mm X 0.109 mm X 0.204 m m L 6 It was mounted along its extended direction. Measurement of unit cell parameters and data collection I I1 proceeded as for H20s3(CO)II. Details are given in Table I. Thesbsorption correction for this crystal was checked by measuring respectively; we assume that replacement of an axial terminal the 331 reflection at x = 90 and 10’ intervals of 4. The observed carbonyl ligand by a hydride ligand will be of little stereovariation in intensity of 37.8% was reduced to 6.4% following correction chemical consequence.] for absorption. Solution of the Structure of H20s3(CO)11. Programs used for Experimental Section structure determination were LSHF (full-matrix least-squares refmement and structure factor calculations, by B. G . DeBoer), FORDM (Fourier Data C o k t i o n for H20s3(C0),,. Yellow crystals of H20s3(CO)Il synthesis, by A. Zalkin), STAN1 (distances and angles, with esd’s, by were provided by Professor J. R. S h q l e y of the University of Illinois DeBoer), and ORTEP (thermal ellipsoid plotting program, by C. K. at Urbana-Champaign. The crystal selected for th_e diffraction Johnson). All calculations were performed on an IBM 370/158 experiment was rod-shaped with ends defined by (101) and (101) computer. separated by 0.213 mm a3d a pentagonal cross section with sides Analytical scattering factors for neutral osmium, carbon, and oxygen defined cyclically by (OOl), (OlO), (021), (OOl), and (010). Perwere taken from the compilation of Cromer and Mann;” Af’and Af” pendicular distances_offa:= from t_he origin (defined by the point were included for all atoms, using the values of Cromer and Libof intersection of (001), (OlO), and (101)) were 0.059 mm for ( e l ) , erman.” 0.087 mm for (OlO), 0.213 mm for (101), and 0.016 mm for (021). The function minimized during least-squares refinement was The crystal was mounted on a glass fiber on a eucentric goniometer ~ W ( ~ # -~ IFcl)2, I where w = 8. Discrepancy indices used below are head. Preliminary photographs yielded approximate cell dimensions, defined as indicated C, Laue symmetry, and revealed the systematic absences
Introduction
Inorganic Chemistry, Vol. 16, No.4, 1977 879
Polynuclear Osmium Carbonyl Hydrides
["""I RF =
'
""3 x
aF,I
100 (%)
Table I. Experimental Data for the X-Ray Diffraction Studies of H20s,(CO),, and Os,(CO),, Compd Temp, "C a, A -
b, A The similar cell dimensions and identical space group of H1O S ~ ( C O and ) ~ ~O S ~ ( C Oled ) ~ us ~ to assume that the species were quasi-isostructural and that the osmium atoms would be located a t approximately the same positions in each species. We therefore took the coordinates of the osmium atoms from Corey and Dahl's d e termination of O S ~ ( C O ) ~along ~ , ' ~ with a scale factor and isotropic thermal parameter from a Wilson plot of the HZO~3(CO)ll data. Refinement of the scale factor and positional and anisotropic thermal parameters for the three osmium atoms converged with R F = 11.3% and RwF = 14.9%. A difference-Fourier map revealed the presence of, apparently, 12 (not 1l!) carbonyl ligands. The carbonyl ligand defined by atoms C(22) and O(22) was the weakest, peak heights being only 2.88 and 2.32 e A-3, respectively. We now included all 12 carbonyl groups in the model, using isotropic thermal parameters for carbon and oxygen atoms. Three cycles of full-matrix least-squares refinement led to convergence with RF = 4.97% and RwF = 5.29%. However, the thermal parameters of C(22) and O(22) were outrageously high,with values of 14.9 and 23.3 Az,respdvely. Following removal of these two atoms from the model, four cycles of refinement led to convergence with little improvement in the discrepancy indices (RF= 4.85% and RwF= 5.27%). A difference-Fourier map revealed peaks of height 2.65 and 2.49 e A-3 at the positions previously occupied by C(22) and O(22). Clearly this ligand was present, but only with partial occupancy. A careful study of thermal parameters for all other carbon and oxygen atoms indicated that the ligand defined by C(21) and O(21) [which is trans to C(22)-0(22)] was also only of partial occupancy. Ligands C(21)-0(21) and C(22)-0(22) were now input with coupled occupancies of a and (1 - a),respectively, with a initially being assigned a value of 0.65. (The ratio of peak heights of C(21)-0(21) vs. C(22)-O(22) was approximately 0.6k0.35.) A structure factor calculation based on this model showed discrepancy indices of RF = 4.49% and R w =~ 4.50% even before refinement, indicating a valid improvement. Continued least-squares refinement with anisotropic thermal parameters for all osmium, oxygen, and carbon atoms and with a secondary extinction correction led to final = 0.014 and with RF= 3.68% and R w ~ convergence with (A/U),,,~~ = 3.52% for all 2259 data (none rejected). The "goodness-of-fit", given by [Cw(lFd - IFd)2/(m- t ~ ) ] 'was / ~ , 0.975 where m,the numhr of observations, was 2259 and n, the number of variables, was 246, yielding an observation-to-variable ratio of 9.18: 1. A final difference-Fourier synthesis had a maximum peak height of 0.88 e A-3. Neither the bridging hydride ligand nor the terminal hydride ligand (vide infra) could be located unambiguously either on this map or on any of a series of other maps which were calculated from data with successively restricted (sin @)/A values. The final value for the secondary extinction coefficient (c) was 5.4 (21) X mm-' 6.where the corrected calculated structure factor
The maximum extinction correction was for the 002 reflection for which Rzc~..,r/~c,u,,~r = 0.923. Refmment of the Structure of OsJCO)Iz. A structural study of Os3(CO)lz was reported by Corey and Dahl in 1962.14 The precision of the study was in keeping with prevailing standards (viz., visual estimation of film data, RF= 12.4% for 1873 data) but not sufficient to allow an accurate comparison of the molecular geometry of ()sj(CO)~z with that of H ~ O S ~ ( C O ) , ~ . Refinement began, using a scale factor and overall isotropic thermal parameter from a Wilson plot of our new data set, in conjunction with the positional coordinates of all atoms from the previous study. Refinement proceeded smoothly; the last few cycles included refinement of the scale factor, positional and anisotropic thermal parameters for all 27 atoms, and a secondary extinction parameter. Convergence ((A/u)- = 0.044)was reached with RF = 3.3596, RwF = 3.62%, and a "goodness-of-fit" of 0.983. The number of o k a t i o n s
c, A
cos p 0, deg
V,A3 Space group
Z
Mol wt p(calcd), g cm-3
(A) Crystal Parameters H,Os,(CO), 21.1 (3) 8.0744 (16) 14.7265 (29) 14.7770 (28) -0.1970 (2) 101.36 (1) 1722.7 (6)
E,/nC 4 880.73 3.396
*
Os,(CO),, 20.8 (3) 8.0817 (6) 14.7683 (11) 14.5773 (11) -0.1832 (1) 100.56 (1) 1710.41 (23) f%/nc
4 906.73 3.521
(B) Measurement of Intensity Data Radiation: Mo Ka Filter(s): Nb foil at counter aperture (-47% transmission of Mo Ka) Attenuators: Cu foil; used i f Z > lo4 countsls Takeoff angle: 3.0" Detector aperture: 3 mm wide (in 20) X 4 mm high (in x) Crystal4etector distance: 330 mm Crystal orientation: for H,Os,(CO),, ,$I axis -4.0" from [ lOi]; for Os3(CO),,, @ axis -1.0" from [22T] Reflections measd: for H,Os,(CO), ,+h,ttk,%lfor 20 < 30" and +h,+k,d for 30" < 20 G 45" ,d for Os,(CO),,, +h,+k,d for 0" < 20 G 50" Scan type: coupled 8(crystal)-20(counter) Scan speed: l.O"/min Scan length: A(20) = (W + 0.692 tan 0)" starting (W/2)" below the Mo Ka, peak;for H,Os,(CO),,, W = 1.00";for Os,(CO),,, W = 0.80" Background measurement: stationary crystal-stationary counter; 20 s each at beginning and end of 20 scan Std reflections: three remeasured after every 47 reflections; rms deviations (after application of an anisotropic linear decay correction)e for H,Os,(CO),, were 0.70% for 21 3, 1.64% for 020, and 0.88%for 204; for Os,(CO),, they were 0.71% for 303,0.76% for 020, and 0.54% for 20T Reflections collected: for HaOs3(CO)11,2259 independent measurements, plus 826 duplicate or equivalent measurements (averaged into primary data set) and 181 systematic absences; for Os,(CO),,, 3041 independent measurements, 18 duplicates, and 139 systematic absences (C) Treatment of Intensity Data Conversione IFoI and u( IF, I): as in ref 15, using "ignorance factors" of p = 0.04 Absorption correction: for H,Os,(CO),, ,p = 221.5 cm-' and the maximum and minimum transmission factorsf were 0.307 and 0.166, respectively; for Os,(CO),,, p = 223.2 cm" and the maximum and minimum transmission factors were 0.196 and 0.1 17, respectively; crystal shapes are described in text a Unit cell parameters are from a least-squares fit to the setting angles of the resolved Mo Ka, peaks of 12 reflections in the 20 range 37-43' and widely dispersed in reciprocal space. The maximum and root-mean-square angular disagreementswere 0.037 and 0.018", respectively ()i(Mo Ka,)0.709 30 A). As in footnote a , except 20 = 41.8-49.1" and the maximum and rms disagreements were 0.019 and 0.008". These values are in agreement with, but more precise than, those reported by Corey and Dahlt4 from film measurements. P2Jn is a nonstandard setting of P2Jc [C2h5; No. 141 having the equipoints+(x,y, z ) and + x, 'I2 - y , '/ ,+ 2 ) . Averaging of the two forms of data (hkl;h, -k, r) collected from the H,Os,(CO),, crystal for 20 < 30" g a v e R I P ) and R(wFz)values of 2.04% and 4.59%, respectively. Prior to application of the absorption correction, these were 4.40% and 6.04%, respectively. e Data reduction and analysis, including decay correction, were performed using the Fortran IV program RDUSP, by B. G. DeBoer. f Absorption corrections and secondary extinction coefficients were calculated by the program DRABZ, by B. G. DeBoer.
was 3040 (none rejected) and the number of parameters refined was 245. The observations-to-parameter ratio was 12.41:l. Secondary extinction was significant, the final value for c (vide mm-' e-2; maximum corrections were supra) being 1.30 ( 5 ) X
880 Inorganic Chemistry, Vol. Id, No. 4, 1977
Melvyn Rowen Churchill and Barry G. DeBoer Table 11. Final Positional Parameters with Esd’f for H20s,(CO),, and Os,(CO),, Atom X Y Z
Figure 1. General view of an H20s3(CO),,molecule with probable positions of hydrogen atoms shown as broken circles (ORTEPdiagram, 30% probability ellipsoids).
Figure 2. Molecular geometry of O S ~ ( C O (ORTEP ) , ~ diagram, 50% ellipsoids). ~ c , c o I / ~ c , u l c o=I
-
0.623 for 002, 0.647 for 204, 0.695 for 113. A final difference-Fourier map showed peaks of height 0.99 0.83 e A-3 and troughs of depth -1.22 -0.92 e k3close to the positions of the osmium atoms. All other features were smaller than these. Positional parameters for H20s3(CO),,and O S ~ ( C Oare ) ~ shown ~ in Table 11; anisotropic thermal parameters are listed in Table 111.
-
Results and Discussion General views of H20s3(CO)11and O S ~ ( C Oappear ) ~ ~ in Figures 1 and 2, respectively. Interatomic distances for the two species are shown Table IV, while interatomic angles are collected in Table V. The structure determined for O S ~ ( C Ois) completely ~~ in accord with that previously published by Corey and Dahl;14 the only difference is that esd’s of parameters in the present determination are reduced to approximately one-tenth of the previous values. The Os3(CO)12 molecule has approximate D3hsymmetry (broken by a minor rotation of the (CO), groups such that equatorial carbonyl ligands are displaced alternately above and below the Os3 plane; see Table VI(B)) with an average Os-Os distance of 2.8771 [27 A.20 The average Os-CO(axia1) bond length is 1.946 [6] , Le., some 0.034 f 0.009 A (or 3 . 8 ~ longer ) than the average Os-CO (equatorial) distance of 1.912 [7] A; carbon-oxygen distances show the reverse pattern, but at a much lower level of significance,with C-O(axia1) = 1.134 [8] A and C-O(equatoria1) = 1.145 [5] A, for a difference of -0.01 1 f 0.009 A. This overall pattern is expected since competition for back-donated electron density is severe for mutually trans carbonyl ligands and close to minimal for carbonyl ligands trans to a simple metal-metal u bond. [Somewhat surprisingly, Mason and Rae21 reported that in R U ~ ( C Othe ) ~Ru-CO(axia1) ~ bonds average 1.89 (2) A as opposed to an average Ru-CO(equatoria1) distance of 1.93 (2) A. However, despite being done fairly recently (1968), this crystallographic analysis is not of the highest precision and the problem could bear reinvestigation.] We now turn our attention to the molecular geometry of H20s3(CO)11.The O S ~ ( C Oportion ) ~ ~ of this molecule has
B
(A) H,Os,(CO),, 0.429 67 (5) -0.039 26 (3) 0.440 89 (6) 0.163 10 (3) 0.689 65 (5) 0.068 01 (3) 0.576 1 (16) -0.045 2 (7) 0.653 0 (13) -0.051 0 (6) 0.289 0 (16) -0.026 2 (9) 0.205 9 (12) -0.021 5 (8) 0.527 7 (16) -0.145 4 (8) 0.588 0 (13) -0.209 9 (7) 0.247 4 (20) -0.092 5 (1 3) 0.133 6 (17) -0.124 6 (11) 0.585 6 (21) 0.162 9 (9) 0.666 4 (16) 0.169 3 (8) 0.280 7 (37) 0.170 7 (14) 0.190 5 (32) 0.180 9 (17) 0.259 6 (16) 0.209 6 (10) 0.152 3 (15) 0.239 3 (9) 0.529 0 (16) 0.274 2 (9) 0.585 2 (14) 0.344 2 (7) 0.827 8 (16) 0.061 l ( 7 ) 0.916 5 (12) 0.059 5 (6) 0.536 1 (17) 0.088 6 (9) 0.456 6 (14) 0.102 2 (8) 0.819 2 (15) 0.168 9 (8) 0.897 4 (13) 0.230 9 (7) 0.812 6 (15) -0.027 9 (8) 0.884 9 (12) -0.084 3 (6)
0.441 82 (4) 0.441 62 (4) 0.696 45 (4) 0.582 5 (12) 0.661 3 (10) 0.302 6 (12) 0.216 6 (9) 0.517 4 (14) 0.569 3 (11) 0.243 2 (13) 0.127 8 (10) 0.584 9 (12) 0.669 2 (11) 0.296 0 (13) 0.205 2 (11) 0.252 3 (13) 0.136 3 (10) 0.517 0 (12) 0.565 6 (11) 0.841 2 (12) 0.931 4 (9) 0.554 2 (11) 0.480 l ( 9 ) 0.824 0 (12) 0.896 5 (10) 0.810 3 (12) 0.879 2 (10)
B,b A’
0.212 95 (3) 3.83 0.200 09 i3j 4.07 3.40 0.330 08 (3) 0.120 0 (10) 4.5 0.065 3 (6) 6.0 5.2 0.306 9 (9) 0.359 2 (8) 7.5 0.271 2 (9) 5.0 0.307 4 (7) 7.7 0.130 6 (11) 7.9 0.083 7 (9) 12.2 0.102 4 (11) 3.1 0.049 7 (8) 4.7 0.294 5 (18) 2.7 0.344 5 (17) 5.8 0.1 10 7 (10) 6.1 0.059 0 (9) 10.7 0.242 2 (9) 5.4 0.272 0 (8) 7.8 0.237 8 (9) 4.8 0.186 5 (7) 6.4 0.416 1 (10) 5.7 0.469 6 (8) 8.1 0.382 8 (8) 4.8 0.412 9 (7) 7.5 0.403 2 (9) 4.9 0.445 0 (7) 6.8
(B) 0~3(CO)j2 -0.024 44 (2) 0.170 11 (2) 0.080 84 (2) -0.031 2 (7) -0.040 4 (5) -0.016 2 (7) -0.016 1 (6) -0.140 8 (7) -0.210 2 (6) -0.052 1 (8) -0.066 8 (7) 0.161 7 (6) 0.163 8 (5) 0.177 5 (7) 0.185 1 (7) 0.185 6 (7) 0.196 2 (7) 0.289 3 (7) 0.361 2 (6) 0.070 7 (6) 0.068 1 (5) 0.097 1 (7) 0.105 7 (5) 0.183 7 (7) 0.244 6 (5) -0.019 3 (7) -0.079 1 (5)
0.222 39 (2) 2.26 0.206 54 (2) 2.37 0.339 17 (2) 2.12 0.127 9 (6) 2.9 0.070 7 (5) 4.3 0.317 6 (6) 2.9 0.370 3 (5) 4.3 0.267 4 (7) 3.5 0.297 6 (6) 5.1 0.129 7 (7) 3.6 0.075 0 (6) 5.5 2.8 0.114 4 (6) 0.058 4 (5) 4.5 0.298 6 (7) 3.3 0.348 7 (6) 5.8 0.108 3 (7) 3.5 0.050 5 (6) 5.5 0.237 4 (7) 3.2 0.257 0 (6) 5.3 0.244 2 (7) 2.7 0.196 l ( 5 ) 3.7 0.431 8 (6) 2.8 0.490 6 (5) 4.3 0.391 6 (6) 3.1 0.424 4 (6) 4.7 0.406 3 (6) 2.9 4.4 0.445 9 (5) a Esd’s (estimated standard deviations) are shown in parentheses and are right-adjusted to the last digit of the preceding number. They are derived from the inverse of the least-squares matrix calculated in the final cycle of refinement. These are the so-called “equivalent” isotropic thermal parameters. They are equal to one-third of the trace of the (orthogonalized) anisotropic thermal tensors. These atoms were assigned fractional occupancies as a model for the disordered carbonyl ligand. Atoms C(21) and O(21) were assigned occupancies of a , and atoms C(22) and 0(22), occupancies of (1 - a ) ;a was least-squares refined to a final value 0f0.638 (11) [ l -a=0.362].
been located directly by x-ray structural analysis. There are two orientations of the molecule in the crystallographic site-one of 63.8% occupancy (shown in Figure 1) and one of 36.2% occupancy, in which the suggested terminal hydrogen site of Figure 1 is replaced by the ligand C(22)-O(22) and
Inorganic Chemistry, Vol. 16, No. 4, 1977 881
Polynuclear Osmium Carbonyl Hydrides Table 111. Final Anisotropic Thermal Parameter@*bfor H,Os,(CO),, and Os,(CO),, Atom
Bl I
4
2
B33
B12
B, 3
(A) HzOs,(CO),i
3.34 (2) 3.77 (2) 3.08 (2) 4.5 (6) 7.2 (6) 3.4 (6) 5.4 ( 5 ) 5.8 (7) 9.1 (7) 5.5 (8) 9.7 (8) 3.4 (8) 5.2 (7) 4.4 (15) 6.8 (15) 3.6 (6) 7.8 (7) 5.6 (7) 8.3 (7) 4.8 (6) 5.6 (5) 4.6 (7) 8.3 (6) 4.6 (6) 8.7 (6) 3.6 (6) 6.3 (5)
3.46 (2) 3.30 (2) 3.26 (2) 3.5 ( 5 ) 6.4 (5) 5.9 (7) 9.8 (7) 3.2 (5) 5.4 (5) 11.5 (12) 18.1 (13) 2.3 (7) 5.5 (7) 0.4 (9) 5.5 (13) 7.8 (9) 12.1 (9) 5.0 (7) 5.0 (5) 4.0 (6) 7.4 (6) 5.5 (7) 9.1 (7) 4.6 (6) 5.8 (5) 4.6 (6) 5.2 (5)
4.34 (2) 4.70 (3) 3.52 (2) 5.0 (7) 4.8 (5) 5.7 (7) 8.2 (6) 6.4 (7) 9.1 (7) 5.5 (8) 7.5 (7) 3.1 (8) 3.7 (6) 2.5 (12) 5.3 (13) 6.6 (8) 11.4 (9) 6.5 (8) 10.2 (7) 5.0 (7) 6.7 (5) 6.8 (8) 8.0 (6) 5.1 (6) 7.5 (6) 6.4 (7) 8.0 (6)
-0.30 (2) 0.72 (2) -0.08 (2) 0.1 (5) -0.4 (4) -1.6 (5) -1.1 (5) 1.2 (5) 1.0 (5) -0.6 (8) -6.1 (9) 0.2 (6) 0.5 (5) 0.4 (9) 0.1 (11) 0.9 (6) 3.7 (6) 0.8 (6) -0.4 (5) -1.9 (5) -0.5 (4) -2.1 (5) -2.3 (5) -0.7 (5) -3.4 (5) -0.8 ( 5 ) 1.0 (4)
-0.09 (2) -0.23 (2) -0.18 (2) -0.2 (5) 1.8 (5) -0.3 (5) 3.2 (5) 2.2 (6) 2.9 (5) -1.8 (7) -1.4 (6) - 0.2 (7) 2.0 (5) -1.0 (11) 1.8 (12) -0.1 (6) -0.4 (6) 3.0 (6) 1.9 (5) -0.3 (6) 2.5 (5) 0.4 (6) 4.4 (6) 0.7 (5) 0.3 (5) 0.4 (5) -1.3 (5)
B23
-0.34 (2) 0.21 (2) 0.03 (2) -0.5 (5) -1.0 (4) -1.7 (6) -2.1 (5) 0.6 (5) 1.7 (5) -0.8 (8) -2.3 (7) -0.2 (6) -0.7 (5) 0.2 (8) -0.9 (10) 2.0 (7) 4.8 (7) 1.1 (6) 0.0 (5) 0.0 (5) -0.1 (4) -1.0 (6) -2.5 (5) 0.3 (5) -1.2 (4) 0.2 (5) 2.5 (4)
0.187, 0.216,0.253 0.181,0.223, 0.268 0.181, 0.203,0.235 0.21, 0.22,0.28 0.23,0.29, 0.31 0.16, 0.27, 0.31 0.22,0.30, 0.38 0.19,0.25,0.30 0.24, 0.31, 0.37 0.19,0.34, 0.38 0.22, 0.38, 0.52 0.17, 0.18,0.23 0.18, 0.26,0.27 0.06, 0.15,0.27 0.23, 0.28,0.30 0.19, 0.28,0.34 0.22,0.39,0.46 0.22, 0.24, 0.32 0.25, 0.32, 0.36 0.17, 0.24,0.30 0.23, 0.30, 0.31 0.19, 0.30, 0.31 0.24, 0.29, 0.41 0.22,0.25, 0.27 0.20, 0.32,0.38 0.20, 0.25,0.29 0.19,0.29,0.37
(B) OS~(CO)IZ 0.155, 0.174, 0.178 -0.132 (11) -0.260 (11) 2.240 (16) 2.279 (16) 0.264 (11) 2.199 (17) 0.154,0.177, 0.187 0.311 (11) 0.303 (11) 2.550 (16) 0.326 (11) 2.243 (17) 2.268 (16) 0.150, 0.168,0.172 -0.094 (11) -0.014 (10) 0.140 (11) 1.921 (15) 2.263 (16) 2.086 (15) 0.16, 0.19,0.22 0.1 (3) 0.0 (3) -0.7 (3) 2.9 (4) 2.6 (4) 3.1 (4) -0.6 (3) 0.20,0.24, 0.26 0.1 (3) 1.8 (3) 4.2 (4) 4.3 (4) 4.9 (4) 0.17, 0.19, 0.22 -0.6 (3) -0.1 (3) 0.2 (3) 2.8 (4) 3.2 (4) 2.6 (4) 0.19, 0.22,0.28 -0.5 (3) -0.9 (3) 1.3 (3) 3.8 (3) 3.5 (3) 5.8 (5) -0.7 (4) -1.1 (4) 0.16, 0.22, 0.24 0.6 (4) 2.8 (5) 3.9 (5) 3.9 (5) 0.18,0.27,0.30 0.6 (3) 3.1 (4) 5.7 (4) 1.1 (3) 0.4 (4) 6.1 (5) -0.6 (4) 0.4 (4) 0.18, 0.20,0.25 2.7 (4) 0.5 (4) 3.5 (4) 4.5 (5) 0.18,0.24,0.35 -1.2 (4) -0.1 (3) -1.8 (4) 4.1 (4) 8.9 (6) 3.3 (4) 0.15,0.18, 0.22 0.5 (3) 0.0 (3) 0.2 (3) 2.7 (4) 3.2 (4) 2.2 (4) 0.20, 0.22, 0.29 1.3 (3) 4.2 (4) 1.0 (3) 1.8 (3) 4.9 (4) 4.9 (4) 0.16, 0.21, 0.23 0.5 (4) 0.9 (4) 0.3 (3) 3.8 (5) 2.6 (4) 3.4 (4) 0.19, 0.25,0.35 2.7 (4) 2.9 (4) 1.3 (4) 5.6 (4) 5.7 (5) 6.7 (5) 0.16, 0.22, 0.24 1.0 (4) 0.4 (4) 0.9 (4) 3.4 (5) 3.3 (5) 3.9 (5) 4.7 (4) 1.2 (4) 0.18, 0.28, 0.32 -0.9 (3) 0.6 (4) 7.5 (5) 3.6 (4) 0.5 (3) 1.0 (3) 0.17,0.20, 0.22 3.2 (4) 3.0 (4) 0.1 (4) 3.5 (5) -1.0 (3) -0.3 (3) 1.4 (4) 0.19,0.28, 0.30 3.2 (4) 6.4 (5) 6.4 (5) -0.1 (3) 0.15,0.16,0.23 1.9 (4) 3.1 (4) -0.5 (3) 0.3 (3) 2.6 (4) 0.16,0.22,0.26 3.4 (3) 1.5 (3) 0.3 (3) 0.3 (3) 2.8 (3) 5.1 (4) 0.14, 0.19, 0.22 2.8 (4) 0.0 (3) 0.2 (3) -0.3 (3) 1.6 (3) 3.8 ( 5 ) 0.19, 0.23, 0.27 -0.5 (3) 3.9 (3) 1.4 (3) -0.2 (3) 3.8 (3) 5.4 (4) 0.0 (4) 0.1 (3) 2.9 (4) 2.7 (4) 0.4 (3) 0.18, 0.19, 0.21 3.5 (5) 0.5 (3) -1.9 (3) -1.9 (3) 0.15, 0.26,0.30 4.0 (4) 4.3 (4) 5.6 (4) -0.3 (3) -0.1 (3) 2.4 (4) 0.17, 0.19,0.21 0.2 (3) 2.8 (4) 3.2 (4) 0.1 (3) 1.1 (3) 1.3 (3) 4.1 (4) 0.17, 0.26,0.27 3.8 (4) 5.0 (4) a These anisotropic thermal parameters are analogous to the usual form of the isotropic thermal parameter and have units of A*. They enter the expression for the structure factor in the form exp[-0.25(Bll h2a*2t B22k2b*2 t B,,E*c*~t 2B, .hka*b* + 2B13hla*c*t 2B2,klb*c*)]. See footnote a to Table 11. These values correspond to the root-mean-square amplitudes of vibration (in A) of the atom along the three principal axes (minor, median, major, respectively) of its vibration ellipsoid. For relative orientations, see figures. See footnote c to Table 11.
in which C(21)-O(21) is absent. Reasonable stereochemistry about each osmium atom can be obtained only by assigning a terminal hydride ligand to the site trans to C(21)-O(21) (as shown in Figure 1) or trans to C(22)-O(22) for the minor orientational component. It is worth noting that the thermal parameters for C(22) are unrealistic-this is probably a result of hydrogen atom contributions not being included in the analysis; C(22) occupies a site which represents the composite of 36.2% occupancy of carbon (approximately 2.2 e) from a carbonyl ligand and 63.8% occupancy of hydrogen (0.6 e) from a hydride ligand, but only the carbon atom contribution was
included in the computation; the observed thermal parameters are therefore reduced from their true values. The bridging hydride ligand can be assigned a site in the equatorial plane, displaced outward from the Os(l)-Os(2) vector and at, or near, the point of intersection of imaginary lines C(24)&(2) and C(13)+0s(l). The effects of this equatorial p2-bridginghydride ligand can be assessed in detail by comparing the geometry of the equatorial O S ~ ( C O )fragments ~ in O S ~ ( C O and ) ~ ~ H2OS~(CO)~ [Note ~ . that there are no major differences in the geometry of the axial ligands of O S ~ ( C O )vs. , ~H20~3(CO)ll.
882 Inorganic Chemistry, Vol. 16, No, 4 , 1977 Table IV. Interatomic Distances (A) with Esd's' for H,Os,(CO),, and Os,(CO),,
Melvyn Rowen Churchill and Barry G. DeBoer Table V. Interatomic Angles (deg) in H,Os,(CO),, and Os,(CO),, Angles in HzOs,(CO)i i
Atoms (A) Osmium-Osmium OS(l)-OS(2) 2.9886 (9) Os(l)-Os(3) 2.9097 (7) Os(2)-0s( 3) 2.8574 (7) Averageb
2.8824 2.8752 2.8737 2.8771
(5) (5) (5) [27]
(B) OsmiumCarbon (Axial) OS(l)C(11) 1.983 (15) 1.943 (10) Os( 1)C(12) 1.968 (15) 1.944 (10) OS(2)4(2 1) 2.028 (18)' 1.931 (10) Os(2)C(22) 2.084 (32)c 1.943 (10) Os( 3 ) C ( 31) 1.926 (15) 1.975 (10) Os(3)C( 32) 1.966 (16) 1.942 (9) Average 1.961 [12] 1.946 [6]
(C) Osmium-Carbon (Equatorial) OS(l)-C(l3) 1.882 (12) 1.900 (11) Os(l)-C(14) 1.886 (16) 1.943 (11) Os(2)C(23) 1.896 (13) 1.909 (10) Os(2)C(24) 1.843 (15) 1.889 (11) Os(3)C (3 3) 1.894 (13) 1.914 (10) Os(3)C(34) 1.930 (13) 1.914 (10) 1.912 [7] Average 1.889 [ l l ] (D) Carbon-Oxygen (Axial) C(11)-0(11) 1.114 (14) 1.146 (11) C(12)-0(12) 1.122 (14) 1.126 (11) C(21)-0(21) 1.114 (18)' 1.156 (11) C(22)-0(22) 1.145 (33)' 1.132 (12) C(31)-0(31) 1.142 (15) 1.101 (12) C( 32)-O( 32) 1.131 (15) 1.140 (11) Average 1.127 [ 7 ] 1.134 [8] (E) Carbon-Oxygen (Equatorial) C(13)-0(13) 1.150 (14) 1.163 (13) C( 14)-0( 14) 1.1 39 (17) 1.131 (12) 1.124 (14) 1.150 (12) C( 2 3 ) 4 (2 3) C(24)-0(24) 1.175 (15) 1.150 (13) C(33)-0(33) 1.150 (13) 1.132 (12) C(34)-0(34) 1.127 (14) 1.142 (11) Average 1.144 [ 8 ] 1.145 [5] (F) Carbon-Carbon (Axial, Nonbonding) C(ll)C(21) 3.078 (18)' 2.856 (13) C(ll)-C(31) 2.867 (15) 2.866 (12) C(2 1 ) C ( 31) 2.922 (18)c 2.871 (12) c(12 ) C(22) 2.906 (25)' 2.874 (14) C( 1 2 ) C (32) 2.860 (15) 2.910 (12) C(22)C(32) 2.735 (26)' 2.837 (13) Average 2.869 [ l o ]
( G ) Oxygen-Oxygen (Axial, Nonbonding) 0(11)-0(21) 3.257 (16)' 3.023 (11) 0(11)-0( 31) 2.979 (12) 3.035 (11) 0(21)-0(31) 3.028 (15)' 2.994 (10) 0(12)-0(22) 2.989 (27)' 2.988 (13) O( 12)-0( 32) 2.963 (14) 3.079 (1 1) O( 22)-0( 32) 2.795 (26)' 2.985 (11) Average 3.017 [15]
'Esd's, shown in parentheses, are right-adjusted to the last digit of the preceding number. Their calculation, using the Fortran IV program STAN1 by B . G . DeBoer, included the effects of uncertainties in unit cell parameters, as well as all correlation coefficients. No corrections have been applied for the effects of thermal Error estimates shown in brackets for average dismotion, tances, d , are the exterior estimates of the precision of the average value given by [ Z n ( d - d ) ' / ( n 2 - n ) ] ' / ' . ' These distances involve atoms of the disordered carbonyl ligand. They were not included in the calculations of average values.
We conclude, therefore, that the terminal hydride ligand in H20s3(CO) has essentially the same stereochemical influence as the corresponding terminal carbonyl ligand in Os3(CO)12 and that all structural dissimilarities between O S ~ ( C Oand ),~ H20s3(CO)11 arise from the presence or absence of a p2bridging hydride ligand in the equatorial plane.]
(A) Within OS(2)-0S( 1)-0s(3) OS(l)-0S(2)-0S(3) 0 s( l)-0s( 3)-0 s(2)
the Os, Triangle 57.93 (2)' 59.65 (1) 62.42 (2)
Angles in Os,(CO)i, 59.88 (1) 59.93 (1) 60.18 (1)
(B) 0 s-0 s-CO(equator ial)
Os(2)4s(l)-C(14) Os(l)-Os(2)C(23) Os(3)-0s(l)C(13) Os( 3)-0s( 2 ) C ( 24) Os(l)-Os( 3 ) C ( 34) 0 ~ ( 2 ) - 03~)(C (3 3)
113.8 (6) 112.0 (5) 89.4 (4) 92.0 (4) 99.6 (4) 98.6 (4)
(C) OC-OsCO(equatoria1) C( 13)-0s( 1)C(14) 98.9 (7) C(2 3)-0~(2)C(24) 96.3 (6) C(33)-0~(3)4(34) 99.4 (5) (D) Os-OsCO(axia1) OS(2)-0S(l)-c(11) 88.3 (3) Os(3)-0s(l)C(ll) 88.7 (3) OS(2)-0S(l)C(12) 88.7 (4) Os( 3)-0s( 1 ) C (12) 88.7 (3) Os( 1)-0s(2)
